The context of this report is the developement of an application able to predict (propose options for) the next word regarding the previous 3 words typed. The purpose of this report is to prepare modeling by an exploratory analysis of the provided datas. These are records of 3 differents sources : blogs, news and twitter. It shows basic summaries about the three data sets provided. These are obtained after tokenizing sentences in the texts, in N-uple of the form word1 word2 word3 .... wordN called N-grams. These N-grams will be used for the prediction. It also gives informations on processing times and memory sizes. Finally, a projection on next steps is detailed.

Select factors for LM

Higher public education institutions in Paris and around

Reproducible Research - Impact of weather events on public health and economy

Reproducible Research: Peer Assessment 1